Agronomy and Horticulture Department


Date of this Version



Supplemental paper for the Proceedings of the Thirty-First North Central Extension- Industry Soil Fertility Conference, Vol. 17, 2001


During the past 35 years, average corn yields in the North Central Region (NCR) have increased at a rate of 1.7 bu/acre per year, mainly due to the adoption of improved crop management technologies and genetic improvement of corn hybrids. Fertilizer K rates used on corn are typically within a range of 0 to 110 lb K2O/acre, but average usage varies widely among states. Commercial fertilizer use rose sharply in the 1960s and 1970s, but corn yield increases since 1980 were achieved with stagnating fertilizer-N use and declining rates of P and K. Signs of emerging K deficiencies have become more common in recent years, particularly in no-till systems. This includes unusual visual symptoms such as K deficiency on younger leaves, but also an unknown range of less visible K deficiencies that are not easily detected based on leaf symptoms. A key question is whether present K management recommendations are adequate to meet future needs. Recent research suggests that (a) commonly used soil tests may not always reflect the actual crop response to K, (b) crop K requirements per unit yield are not constant, but vary with the absolute yield levels and crop management factors, (c) spatial variability of soil K affects K management strategies, (d) genotypic differences exist in the response to soil and fertilizer K, and (e) non-yield traits such as stalk strength or product quality must be taken into account in K management decisions. Therefore, future, fertilizer recommendation algorithms should be more robust and accommodate different crops, cropping systems, crop management technologies, soil conditions, and climate-driven yield potential. Such refinements can be made at different levels of complexity such that a general recommendation can be broken down into more specific recommendations. Agroecological zoning and crop simulation models should play a major role in making these refinements.